We are interested in the biophysical properties and structure of native chromatin fragments. The chicken folate receptor and beta-globin gene loci are ideal for such structural studies in that (i) the region possesses both condensed and transcriptionally active chromatin regions, and (ii) the system has been extensively studied in terms of gene regulation, allowing us to relate the overall chromatin structure to transcription. The condensed chromatin region, spanning 16 kbp of DNA flanked by the developmentally regulated folate receptor and beta-globin genes, can be released from the cell nucleus with the restriction enzyme HpaII. We have previously combined biophysical methods with real-time PCR and Southern blotting to analyze the hydrodynamic properties of this 16 kbp condensed chromatin and showed that it is an extended rod. This provides insights into the structure of heterochromatin, found interspersed within various genomes. We have furthered our studies of native chromatin fragments utilizing the methods we developed. Using an erythroid precursor cell line, we have analyzed a 16 kbp region of the transcriptionally poised beta-globin gene locus and showed that this chromatin fragment also appears to be an extended rod. However, unlike the condensed chromatin fragment, this chromatin region has a lower protein to nucleic acid ratio suggesting a role for RNA in the stabilization of this complex. Similar studies on 10-day old chicken erythrocyte chromatin, in which the beta-globin locus is transcriptionally active, are in progress. The persistence length (flexibility) of the chromatin fragments studied provides additional information and we are currently developing biochemical methods to determine the flexibility of such globin gene chromatin fragments. Macromolecular assemblies. In collaboration with members of the Laboratory of Molecular Biology, and other laboratories, protein and protein-nucleic acid assemblies have been characterized in terms of their shape, stoichiometry and affinity of interaction using hydrodynamic methods. These methods complement biochemical and structural investigations and provide important information on the biological mechanism. A case in point is provided by the recent studies of the conserved MutL protein carried out with Dr. Wei Yang. We showed that the C-terminal domain of MutL is a dimer. Furthermore, we showed that the proline-rich linker connecting the N- and C-termini is essentially rigid resulting in a greater than 10 nm end-to-end distance for the rod-shaped protein. Together with other data, these studies provide a working model for DNA mismatch repair in E. coli.